Hey Team— For the TiDE and other neuralforecast model, is there a way to show the importance of each of the exog features passed into the model? Thanks!

No, I'd suggest to use KernelExplainer from Shap. It's slow though.